Hi,
When I use the new restart job function, it will relaunch the job directly, not respecting the current queue. So if there are jobs ahead of it in the queue, both it and the job ahead of it will be submitted to the same GPU. Is this the intended behavior?
Cheers
Oli